A method for managing audio and video information based on data attributes

Abstract

The present invention relates to the field of data processing technology, and specifically to a method for managing audio and video information based on data attributes, comprising: acquiring video data, constructing a cloud database, storing the video data using the cloud database, marking the video data source user ID with the corresponding video data during the storage of the video data in the cloud database, and then performing the storage operation of the video data; the present invention provides a method for providing specified attribute parameters in the video data and constructing a priori database to provide a more effective and more accurate audit service for the video data, thereby ensuring that short videos containing unhealthy information can be screened and processed, effectively reducing the workload of video data audit staff, and at the same time, based on the health analysis of the short video, further performing a digital assessment of the video data source user, and further assisting the short video APP background management user to carry out security management of the short video APP users.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a method for managing audio and video information based on data attributes. Background Art

[0002] Short video review is a critical component in ensuring the integrity of online content. Through the collaborative work of a professional team and intelligent algorithms, we comprehensively review video content. We accurately identify unhealthy content and analyze it from multiple perspectives, including image, audio, and subtitles, to ensure that published short videos comply with laws, regulations, and platform standards, fostering a positive and safe content environment for users.

[0003] The invention patent application with application number 202410064265.1 discloses a short video review method, including: obtaining video data; labeling the video data according to video quality and evaluation indicators to obtain labeled data; constructing prompt data according to the video data based on the evaluation indicators; and training a large model based on the labeled data and the prompt data to obtain a trained review model, so as to use the review model to review short videos; wherein, the evaluation indicators are used to evaluate the difficulty characteristics, emotional characteristics and highlight characteristics of the video data.

[0004] The application aims to address the problem that "with the explosive growth in the number of short videos, the quality of content is uneven. This not only reduces the overall viewing experience, but may also cause user dissatisfaction and even damage the reputation of the platform."

[0005] However, the video data uploaded by users on short video platforms are still mainly reviewed manually. The existing review technology can often only achieve a rough preliminary review effect as an auxiliary tool, and the review effect is poor.

[0006] To this end, a governance method for audio and video information based on data attributes is proposed. Summary of the Invention

[0007] In view of the above-mentioned shortcomings of the prior art, the present invention provides a method for managing audio and video information based on data attributes, which solves the technical problems raised in the above-mentioned background technology.

[0008] To achieve the above objectives, the present invention is implemented through the following technical solutions:

[0009] A method for managing audio and video information based on data attributes, comprising:

[0010] Obtain video data, build a cloud database, and use the cloud database to store video data. When the video data is stored in the cloud database, mark the video data source user ID with the corresponding video data, and then perform the video data storage operation; traverse the video data stored in the cloud database, extract video data attributes, analyze the video data health based on the video data attributes, and evaluate the safety index of the video data source user according to the video data health analysis results; set a video data health judgment threshold, set a video data source user security judgment threshold, and compare the set judgment threshold with the video data health analysis results and the video data source user security index evaluation results; decide whether to agree to the user to publish the video data based on the comparison results, and decide the user's operation permissions in the short video APP based on the comparison results; obtain the unsafe video data source user based on the comparison results, identify the storage location of the video data published by the video data source user in the cloud database, and sniff out the security risk user according to the identification results; push prompt information to the security risk user when the security risk user comes online.

[0011] Furthermore, the video data is derived from a user group that has registered and logged into the short video APP. During the video data storage phase, the cloud database synchronously obtains the associated users of each user in the user group from the short video APP terminal backend. The cloud database differentiates each user in the user group based on the associated user acquisition results, and synchronously sets a number of differentiated storage intervals. The differentiated storage intervals are used to store video data posted by associated users, so that each video data stored in each differentiated storage interval has at least one common associated user.

[0012] Among them, users in the user group of unrelated users occupy an independent partitioned storage interval in the cloud database. Every time the associated users of the user group change, the video data stored in each partitioned storage interval in the cloud database will be changed and migrated synchronously.

[0013] Furthermore, the attributes of the video data include: text information recognized in the video data, the maximum percentage of pixels belonging to the skin color pixel value interval in each frame of the video data, the trajectory of dynamic target body movements in the video data, and the recognition results of object names in the scene of the video data;

[0014] Build a database for storing unhealthy text information, unhealthy body movement trajectories, and names of objects that should not appear in video data, set pixel value intervals representing skin color, and analyze the healthiness of video data based on video data attributes;

[0015] ;

[0016] Where: K is the health performance value of video data; q(pin ∩p s ) is the number of intersections between the text information identified in the video data and the unhealthy text information stored in the database; q(p s ) is the number of words and characters storing unhealthy text information in the database; k(max) is the maximum ratio of the number of pixels belonging to the skin color pixel value interval in each frame of the video data; n and m are the total number of dynamic target limb movement trajectories in the video data obtained by dividing the video data based on the specified time interval and the total number of unhealthy limb movement trajectories stored in the database; SIMM(L i ,L j ) is the similarity between the i-th action trajectory and the j-th action trajectory; q(C in ∩C s ) is the intersection of the total number of object names recognized in the scene where the video data is located and the total number of object names stored in the database that should not appear in the video data; q(C s ) is the total number of object names stored in the database that should not appear in video data; ω1, ω2, ω3, ω4 are normalized weights;

[0017] in, Express The average operation of ω1, ω2, ω3, and ω4 are all between 0 and 1. The larger the video data health performance value K is, the unhealthier the video data is, and vice versa.

[0018] Furthermore, the motion trajectory is represented by a three-dimensional multi-segment broken line obtained by connecting multiple three-dimensional spatial coordinates in a three-dimensional space, and the similarity of the motion trajectory is calculated using the following formula:

[0019] ;

[0020] Where: SIMM(L a ,L b ) is the similarity of the two motion trajectories; is the ratio of the lengths of the two trajectories; u is the total number of nodes on the trajectory except the first and last points; is the ratio of the expansion angles of the line segments connecting the two trajectories at the vth node on the two trajectories; is the ratio of the total length of the line segment connecting the two motion trajectories at the vth node on the two motion trajectories; ε is the adjustment factor;

[0021] Among them, the adjustment factor ε takes the value of 1 or -1. If the numerator of the fraction where the adjustment factor is located is greater than the denominator, the adjustment factor ε takes the value of -1. If the numerator of the fraction where the adjustment factor is located is less than or equal to the denominator, the adjustment factor ε takes the value of 1. Express The averaging operation.

[0022] Furthermore, the safety index evaluation logic of the video data source user is expressed as:

[0023] All video data released by the video data source users are evaluated for healthiness performance values, and the results are sorted by the release time of the corresponding video data and recorded as K1, K2, K3, ...;

[0024] ;

[0025] Where: K saf is the security index of the user who provided the video data; max(K1, K2, K3, ...) is the maximum value in the brackets; min(K1, K2, K3, ...) is the minimum value in the brackets;

[0026] Among them, the security index K of the user from whom the video data comes saf The larger the value, the lower the security of the user who provided the video data. Conversely, the smaller the value, the higher the security of the user who provided the video data.

[0027] Furthermore, the video data health determination threshold and the video data source user safety determination threshold are both customized by the user end;

[0028] The configuration of user operation permissions in the short video app and the user security index assessment results is subject to the following:

[0029] ;

[0030] Where: K norr is the user security determination threshold for the video data source; [K1, K2), [K2, K3), [K3, K4), ... are the permission configuration levels;

[0031] Among them, [K1, K2), [K2, K3), [K3, K4), ... are customized by the user end, and each interval in [K1, K2), [K2, K3), [K3, K4), ... corresponds to different operation permission content.

[0032] Furthermore, the operation of identifying the storage location of the video data released by the video data source user in the cloud database is the operation of identifying the corresponding differentiated storage intervals of the video data source user in the cloud database, and the operation of sniffing out the security risk users is the operation of sniffing out the storage of each video data source user in the differentiated storage interval, and all sniffed video data source users are recorded as security risk users.

[0033] Furthermore, the user with security risks goes online in the short video APP, and the prompt information pushed to the user with security risks is text information, and the content of the prompt information is manually edited by the short video APP terminal management user.

[0034] Furthermore, when deciding the user's operating permissions in the short video app based on the comparison results, the security index assessment results of the video data source user are continuously recorded. The security index assessment results of the continuously accumulated video data source users are presented in the form of a line graph based on the time series. The corresponding users whose line graph shows a continuous increase for x consecutive times are determined to be high-risk users;

[0035] Among them, x is customized by the short video APP terminal background, and the short video APP terminal background performs a ban operation on high-risk users.

[0036] Compared with the known public technology, the technical solution provided by the present invention has the following beneficial effects:

[0037] The present invention provides a method for managing audio and video information based on data attributes. During the execution of the method, by providing specified attribute parameters in the video data and constructing a priori database, the method provides a more effective and accurate audit service for the video data, ensuring that short videos containing unhealthy information can be screened and processed, effectively reducing the workload of video data audit staff, and based on the health analysis of short videos, further digitally evaluating the source users of the video data, further assisting the short video APP background management users to carry out security management of the short video APP users. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] To more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. Those skilled in the art can also derive other drawings based on these drawings without inventive effort.

[0039] Figure 1 The present invention is a flowchart of a method for managing audio and video information based on data attributes. DETAILED DESCRIPTION

[0040] To make the purpose, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making any creative efforts shall fall within the scope of protection of the present invention.

[0041] The present invention will be further described below with reference to the embodiments.

[0042] Example:

[0043] This embodiment is a method for managing audio and video information based on data attributes, such as Figure 1 As shown, including:

[0044] Acquire video data, build a cloud database, and use the cloud database to store the video data. When the video data is stored in the cloud database, tag the video data source user ID to the corresponding video data, and then perform the video data storage operation;

[0045] The video data comes from the user group that registers and logs in to the short video APP. When the cloud database is used for video data storage, it synchronizes with the short video APP terminal backend to obtain the associated users of each user in the user group. Based on the associated user acquisition results of each user in the user group, the cloud database distinguishes each user in the user group and simultaneously sets several groups of differentiated storage intervals. The differentiated storage intervals are used to store the video data released by the associated users, so that the corresponding users of the video data stored in each differentiated storage interval have at least one common associated user.

[0046] Among them, users in the user group of unrelated users occupy an independent partitioned storage interval in the cloud database. Every time the associated users of the user group change, the video data stored in each partitioned storage interval in the cloud database will be changed and migrated synchronously;

[0047] Traverse the video data stored in the cloud database, extract the video data attributes, analyze the health of the video data based on the video data attributes, and evaluate the security index of the user who provided the video data based on the video data health analysis results;

[0048] The attributes of the video data include: text information recognized in the video data, the maximum percentage of pixels in the skin color pixel value range in each frame of the video data, the trajectory of the dynamic target's body movements in the video data, and the recognition results of the names of objects in the scene of the video data;

[0049] Build a database for storing unhealthy text information, unhealthy body movement trajectories, and names of objects that should not appear in video data, set pixel value intervals representing skin color, and analyze the healthiness of video data based on video data attributes;

[0050] ;

[0051] Where: K is the health performance value of video data; q(p in ∩p s) is the number of intersections between the text information identified in the video data and the unhealthy text information stored in the database; q(p s ) is the number of words and characters storing unhealthy text information in the database; k(max) is the maximum ratio of the number of pixels belonging to the skin color pixel value interval in each frame of the video data; n and m are the total number of dynamic target limb movement trajectories in the video data obtained by dividing the video data based on the specified time interval and the total number of unhealthy limb movement trajectories stored in the database; SIMM(L i ,L j ) is the similarity between the i-th action trajectory and the j-th action trajectory; q(C in ∩C s ) is the intersection of the total number of object names recognized in the scene where the video data is located and the total number of object names stored in the database that should not appear in the video data; q(C s ) is the total number of object names stored in the database that should not appear in the video data; ω1, ω2, ω3, ω4 are normalized weights;

[0052] in, Express The average operation of ω1, ω2, ω3, and ω4 are all between 0 and 1. The larger the video data health performance value K is, the less healthy the video data is, and vice versa.

[0053] The health performance value of the video data is calculated through the above logical formula, providing necessary computational data support for the subsequent calculation of the safety index of the user from whom the video data is sourced by the method in this embodiment.

[0054] The motion trajectory is represented by a three-dimensional multi-segment polyline obtained by connecting multiple three-dimensional spatial coordinates in a three-dimensional space. The similarity of the motion trajectory is calculated using the following formula:

[0055] ;

[0056] Where: SIMM(L a ,L b ) is the similarity of the two motion trajectories; is the ratio of the lengths of the two trajectories; u is the total number of nodes on the trajectory except the first and last points; is the ratio of the expansion angles of the line segments connecting the two trajectories at the vth node on the two trajectories; is the ratio of the total length of the line segment connecting the two motion trajectories at the vth node on the two motion trajectories; ε is the adjustment factor;

[0057] Among them, the adjustment factor ε takes the value of 1 or -1. If the numerator of the fraction where the adjustment factor is located is greater than the denominator, the adjustment factor ε takes the value of -1. If the numerator of the fraction where the adjustment factor is located is less than or equal to the denominator, the adjustment factor ε takes the value of 1. Express The mean operation of ;

[0058] The above logic formula further defines the logic for calculating the similarity between motion trajectories.

[0059] The security index evaluation logic of the video data source user is expressed as follows:

[0060] All video data released by the video data source users are evaluated for healthiness performance values, and the results are sorted by the release time of the corresponding video data and recorded as K1, K2, K3, ...;

[0061] ;

[0062] Where: K saf is the security index of the user who provided the video data; max(K1, K2, K3, ...) is the maximum value in the brackets; min(K1, K2, K3, ...) is the minimum value in the brackets;

[0063] Among them, the security index K of the user from whom the video data comes saf The larger the value, the lower the security of the user who provided the video data. Conversely, the smaller the value, the higher the security of the user who provided the video data.

[0064] The security index of the user who sources the video data is calculated using the above logical formula.

[0065] Set a threshold for judging the health of video data and a threshold for judging the security of the user who provided the video data. Compare the thresholds with the results of the video data health analysis and the security index assessment results of the user who provided the video data.

[0066] The video data health judgment threshold and the video data source user safety judgment threshold are both customized by the user end;

[0067] The configuration of user operation permissions in the short video app and the user security index assessment results is subject to the following:

[0068] ;

[0069] Where: K norr is the user security determination threshold for the video data source; [K1, K2), [K2, K3), [K3, K4), ... are the permission configuration levels;

[0070] Among them, [K1, K2), [K2, K3), [K3, K4), ... are customized by the user end, and each interval in [K1, K2), [K2, K3), [K3, K4), ... corresponds to different operation permissions;

[0071] Through the above logical formula, the logic for the mutual configuration of user operation permissions in the short video APP and user security index assessment results is limited.

[0072] Based on the comparison results, decide whether to allow the user to publish the video data, and based on the comparison results, decide the user's operation permissions in the short video app;

[0073] Based on the comparison results, the source user of the unsafe video data is obtained, the storage location of the video data released by the source user of the video data in the cloud database is identified, and the user with security risks is sniffed out based on the identification results;

[0074] Identify the storage location of the video data released by the video data source user in the cloud database, that is, identify the corresponding storage interval of the video data source user in the cloud database, and sniff the operation of the security risk user, that is, sniff the operation of each video data source user stored in the storage interval. All sniffed video data source users are recorded as security risk users;

[0075] When a user with security risks comes online, push a reminder message to the user with security risks;

[0076] When a user with security risks goes online in the short video APP, the prompt information pushed to the user with security risks is a text message, and the content of the prompt information is manually edited by the short video APP terminal management user.

[0077] In this embodiment, by executing the method in the above embodiment, a better review process is provided for the short videos uploaded to the short video platform, ensuring that the video quality for user groups to interact on the short video platform is better, which is conducive to promoting a healthier and more harmonious environment for the short video platform.

[0078] like Figure 1 As shown, according to the comparison results, when deciding the user's operation permission in the short video APP, the security index assessment results of the video data source user are continuously recorded, and the security index assessment results of the continuously accumulated video data source user are represented in the form of a line graph based on the time series. The corresponding users whose line graph shows a continuous increase for x consecutive times are determined to be high-risk users;

[0079] Among them, x is customized by the short video APP terminal background, and the short video APP terminal background performs a ban operation on high-risk users.

[0080] Through the above settings, the method in the above embodiment is used to further monitor the user's safety index evaluation results, thereby facilitating the management of short video APP users by the short video APP management background.

[0081] In summary, during the execution of the method in the above embodiment, by providing specified attribute parameters in the video data and constructing a priori database, it provides a better and more accurate review service for the video data, ensuring that short videos containing unhealthy information can be screened and processed, effectively reducing the workload of video data review staff, and at the same time, based on the health analysis of short videos, further digital evaluation of the source users of the video data is carried out, further assisting the short video APP background management users to carry out security management of the users using the short video APP.

[0082] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements will not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A method for managing audio and video information based on data attributes, characterized in that: include: Acquire video data, build a cloud database, and use the cloud database to store the video data. When the video data is stored in the cloud database, tag the video data source user ID to the corresponding video data, and then perform the video data storage operation; Traverse the video data stored in the cloud database, extract the video data attributes, analyze the health of the video data based on the video data attributes, and evaluate the security index of the user who provided the video data based on the video data health analysis results; The attributes of the video data include: text information recognized in the video data, the maximum percentage of pixels belonging to the skin color pixel value interval in each frame of the video data, the trajectory of dynamic target body movements in the video data, and the recognition results of object names in the scene of the video data; Build a database for storing unhealthy text information, unhealthy body movement trajectories, and names of objects that should not appear in video data, set pixel value intervals representing skin color, and analyze the healthiness of video data based on video data attributes; ； Where: K is the health performance value of video data; q(p in ∩p s ) is the number of intersections between the text information identified in the video data and the unhealthy text information stored in the database; q(p s ) is the number of words and characters storing unhealthy text information in the database; k(max) is the maximum ratio of the number of pixels belonging to the skin color pixel value interval in each frame of the video data; n and m are the total number of dynamic target limb movement trajectories in the video data obtained by dividing the video data based on the specified time interval and the total number of unhealthy limb movement trajectories stored in the database; SIMM(L i ,L j ) is the similarity between the i-th action trajectory and the j-th action trajectory; q(C in ∩C s ) is the intersection of the total number of object names recognized in the scene where the video data is located and the total number of object names stored in the database that should not appear in the video data; q(C s ) is the total number of object names stored in the database that should not appear in the video data; ω1, ω2, ω3, ω4 are normalized weights; in, Express The average operation of ω1, ω2, ω3, and ω4 are all between 0 and 1. The larger the video data health performance value K is, the less healthy the video data is, and vice versa. Set a threshold for judging the health of video data and a threshold for judging the security of the user who provided the video data. Compare the thresholds with the results of the video data health analysis and the security index assessment results of the user who provided the video data. Based on the comparison results, decide whether to allow the user to publish the video data, and based on the comparison results, decide the user's operation permissions in the short video app; Based on the comparison results, the source user of the unsafe video data is obtained, the storage location of the video data released by the source user of the video data in the cloud database is identified, and the user with security risks is sniffed out based on the identification results; When a user with security risks comes online, a prompt message is pushed to the user with security risks.

2. A method for managing audio and video information based on data attributes according to claim 1, characterized in that: The video data is derived from a user group that has registered and logged into the short video APP. During the video data storage phase, the cloud database synchronizes with the short video APP terminal backend to obtain the associated users of each user in the user group. Based on the associated user acquisition results of each user in the user group, the cloud database distinguishes each user in the user group and simultaneously sets a number of differentiated storage intervals. The differentiated storage intervals are used to store video data posted by associated users, so that each video data stored in each differentiated storage interval has at least one common associated user. Among them, users in the user group of unrelated users occupy an independent partitioned storage interval in the cloud database. Every time the associated users of the user group change, the video data stored in each partitioned storage interval in the cloud database will be changed and migrated synchronously.

3. The method for managing audio and video information based on data attributes according to claim 1, characterized in that: The motion trajectory is represented by a three-dimensional multi-segment broken line obtained by connecting multiple three-dimensional spatial coordinates in a three-dimensional space. The similarity of the motion trajectory is calculated using the following formula: ； Where: SIMM(L a ,L b ) is the similarity of the two motion trajectories; is the ratio of the lengths of the two trajectories; u is the total number of nodes on the trajectory except the first and last points; is the ratio of the expansion angles of the line segments connecting the two trajectories at the vth node on the two trajectories; is the ratio of the total length of the line segment connecting the two motion trajectories at the vth node on the two motion trajectories; ε is the adjustment factor; Among them, the adjustment factor ε takes the value of 1 or -1. If the numerator of the fraction where the adjustment factor is located is greater than the denominator, the adjustment factor ε takes the value of -1. If the numerator of the fraction where the adjustment factor is located is less than or equal to the denominator, the adjustment factor ε takes the value of 1. Express The averaging operation.

4. The method for managing audio and video information based on data attributes according to claim 1, characterized in that: The safety index evaluation logic of the video data source user is expressed as follows: All video data released by the video data source users are evaluated for healthiness performance values, and the results are sorted by the release time of the corresponding video data and recorded as K1, K2, K3, ...; ； Where: K saf is the security index of the user who provided the video data; max(K1, K2, K3, ...) is the maximum value in the brackets; min(K1, K2, K3, ...) is the minimum value in the brackets; Among them, the security index K of the user from whom the video data comes saf The larger the value, the lower the security of the user who provided the video data. Conversely, the smaller the value, the higher the security of the user who provided the video data.

5. The method for managing audio and video information based on data attributes according to claim 1, characterized in that: The video data health determination threshold and the video data source user safety determination threshold are both customized by the user end; The configuration of user operation permissions in the short video app and the user security index assessment results is subject to the following: ； Where: K norr is the user security determination threshold for the video data source; [K1, K2), [K2, K3), [K3, K4), ... are the permission configuration levels; Among them, [K1, K2), [K2, K3), [K3, K4), ... are customized by the user end, and each interval in [K1, K2), [K2, K3), [K3, K4), ... corresponds to different operation permission content.

6. The method for managing audio and video information based on data attributes according to claim 1, characterized in that: The operation of identifying the storage location of the video data released by the video data source user in the cloud database is the operation of identifying the corresponding distinguished storage interval of the video data source user in the cloud database. The operation of sniffing the security risk user is the operation of sniffing the storage of each video data source user in the distinguished storage interval. All sniffed video data source users are recorded as security risk users.

7. The method for managing audio and video information based on data attributes according to claim 1, characterized in that: The user with security risks goes online in the short video APP, and the prompt information pushed to the user with security risks is text information, and the content of the prompt information is manually edited by the short video APP terminal management user.

8. The method for managing audio and video information based on data attributes according to claim 1, characterized in that: Based on the comparison results, the user's operation permission stage in the short video app is determined. The security index assessment results of the video data source users are continuously recorded. The security index assessment results of the continuously accumulated video data source users are presented in the form of a line graph based on the time series. The corresponding users whose line graph shows a continuous increase for x consecutive times are identified as high-risk users. Among them, x is customized by the short video APP terminal background, and the short video APP terminal background performs a ban operation on high-risk users.

Images